The spatial distribution of meteorological elements is important for understanding the regional meteorology and climate changes. However, previous studies rarely focused on the daily changes of the spatial patterns of meteorological elements due to the limitation of remote sensing (RS) techniques and traditional meteorological methods. In this paper, the regional meteorological elements were simulated by the fifth-generation non-hydrostatic Mesoscale Model (MM5), and the spatial patterns of meteorological elements and their diurnal variations were analyzed in landscape level over the Pearl (Zhujiang) River Delta (PRD), China. The results showed that there were several centers of urban heat islands, cold islands, dry islands, wet islands, high wind over the PRD at noon. The diurnal changes of Moran I of meteorological elements were obvious and they reached the extremum at noon and 2-3 hours after the sunrise. The landscape indices of meteorological elements, such as area-weighted mean Fractal Dimension Index (FRAC_AM), Landscape Shape Index (LSI), Shannons Diversity Index (SHDI) and Contagion Index (CONTAG), were more variable at about the sunrise, noon and sunset. The occurrence of wave crests and vales of landscape indices was affected by the surface net radiation, turbulence and local circumfluence. The spatial patterns of meteorological elements correlated well with the land surface, thermal exchanges and local circumfluence. A new approach combining GIS, RS and numerical simulatious technologies and the landscape ecology method was presented to analyze spatial patterns of meteorological elements, which may be useful for studying global and regional climate changes.